Manipulation of the Mitsubishi robot RV-2 with an artificial vision

Abstract

This article presents a methodology for the manipulation of the Mitsubishi RV-2A robot. The manipulation involves the design and implementation of software for the selection of events of visual means previously scheduled with trajectories on the industrial robot through eye movement of a determined person. Capturing eye movements includes the use of a filter to manage the image after and using artificial intelligence techniques and then training a neural network for the recognition of the characteristics of the captured image and its position and with this information the software activates the control system of the robot to position it. Tests were carried out experimentally using five different objects 15 cm apart, noting satisfactory answers